Research On The Influence Of Consignors' Behavior Preference On User Value In NTOCC Platform

This research takes the consignors of NTOCC(Non-Truck Operating Common Carrier)Platform as research objects,and analysis the influence of consignors' behavior preference on user value.Firstly,the paper expounds the development status of the NTOCC platform,points out the operation difficulties faced at present.According to the similarity and particularity between the NTOCC platform and other electronic business platform,it improves the traditional user value model,constructs the RFMD user value model which is applicable to the NTOCC platform,and formulates the value measurement standard for the consignors.The study takes consignors' transaction data of Y company as an example,uses K-Means algorithm to subdivide the consignors,and divides them into four categories: the general maintenance users,the important value users,the important recall users and the general retention users.Based on the clustering results,the four categories,features are selected by Boruta algorithm and PCA algorithm respectively.Comparing the results of this two algorithms,it is found that almost all the features are listed as important features by the Boruta algorithm,obviously the reference value of this result is not great.While the PCA dimension reduction selects 5-6 significant features for each category,and the result is effective.then,the features selected by PCA algorithm has been obtained,and which has been used to describe the four categories of the consignors.Finally,on the basis of the features of different categories,it puts forward practical and accurate operation suggestions for Y company,which provides industry operation guidance for the NTOVV platform.